[jstevewhite]

pentaxq-compare: godfrog: Galleries: Digital Photography Review
I wonder where they got the Q from, but this doesn't look too good. The NEX-5 obviously just kills the Q, the S95 looks a bit better too, though the difference isn't as big. The S95 is a bit more naturalistic to me, and captures more detail. And of course both cameras are cheaper than the Q.

Oh, btw., seems like dpreview got their hands on a Q.
Pentax Q Compact System Camera Samples - Hands-On Preview The photos are gone, but they may be the source of the above photo.

I would expect the NEX to do better, particularly at ISO3200, with its MUCH larger sensor. Frankly, I think it's (the Q) got astonishingly high quality for an ISO3200 image from such a small sensor, if this is really at ISO3200. I disagree that the S95 looks better in the image you linked to. Lines are clearly broken up by smeary noise reduction on the S95, while the Q shot looks cleaner and has better color (more like the NEX, anyway).

When does the Q ship? I'm curious to see real images.

[froeschle]

Here's a concept that's missing from your equivalence equation: Final magnification. Final magnification may be the single highest determiner of IQ of those discussed here so far, and your equivalence model ignores it completely.

Display size

On the Super-Duper Quick: Equivalence is a framework for comparing the IQ of different formats on the basis of five parameters (perspective, framing, DOF, shutter speed, and display size), which all directly relate to the visual properties of the final photo, and are independent of the technology. In addition, the same amount of light will fall on the sensor for Equivalent photos, which will result in the same image noise for equally efficient sensors.

I think - as "fast" and "slow" are rather ill-defined in this terminology - it is not really worth arguing about.

No, it won't. It uses a different sensor and different electronics and a different lens and all of these things add up to differences that will not be captured by "equivalence". You even said as much: "The "equivalence" considerations are based on an "everything else being equal" assumption. In practice, "everything else" is rarely the same. Sensors differ in their noise floors, processing uses different levels of (hidden) noise reduction, etc."

Equivalence is the best one can do with simple calculations. Setting the same f-stop for different sensor sizes and using identical ISO is much worse.
Exposure paradigm

As mentioned in the introduction of this essay, the concept of Equivalence is controversial because it replaces the paradigm of exposure, and its agent, f-ratio, with a new paradigm of total light, and its agent, aperture.

[jstevewhite]

Display size

I think - as "fast" and "slow" are rather ill-defined in this terminology - it is not really worth arguing about.

Equivalence is the best one can do with simple calculations. Setting the same f-stop for different sensor sizes and using identical ISO is much worse.
Exposure paradigm

Obviously Class A disagrees, since his assertion is that the f1.9 lens is slow and that's what causes the image quality difference, not the small sensor, when in fact, all things being equal, final magnification *is* the problem (the small sensor). And obviously I disagree since I think such assertions are misleading.

The entire discussion, when presented as it was, is confusing to people who don't need it. The bare assertions were such that people reading it believed the claim was that I would have to use a different exposure to take a picture. I understand "equivalence" as expressed; I think it's not very valuable. Attempting to re-define exposure *doesn't help anyone* understand photography.

And there are a couple of counter-factual claims at issue here: That different focal length lenses let through different amounts of light as a result of the physical size of their aperture - that is, that a 28mm lens @f1.9 passes less light than a 200mm lens @f1.9.

Furthermore, I suspect that all of this discussion aside, an 8x10 from the Q and an 8x10 from the K20D, shot in bright sunlight, with equivalent focal lengths and apertures (say, f8, normal lens, focussed on a group of folks twenty feet away), would be virtually indistinguishable - that is, you'd have to see 'em side-by-side to tell the difference, and then you'd have to look for cues like DOF and close edge detail, and if you weren't *very* technical, you wouldn't be able to pick a "winner" (edit) - and at 'web sizes', I'm betting they'll be completely indistinguishable in many applications.

Equivalence doesn't provide "real world information". I read Mr. James' essay, and I think it's much ado over nothing, and makes assumptions I disagree with, such as the ones surrounding "display size" he discusses.

[rparmar]

Hey Laurentiu, I see now why you find my statements too generic. I think all statements work that way once taken out of context.

You say "I wouldn't have to keep pointing the logical errors that you are making" which implies I am making more than one, on a continuous basis. And yet the only possible example you can point to is a single instance of a loaded question.

You say that "Loaded questions are designed to fool people" but I was not trying to fool anyone -- I completely stand by my assertions. Which is to say that brand bias affects people's judgement of products like the Pentax Q. So I guess now I have made this point of view crystal clear you must acknowledge your mistake: I am not trying to "fool" anyone. But obviously my observation pushed your button.

Finally, your claim that my sentence "is as offensive as you can get in a written conversation" merely demonstrates that you haven't read much on the internet. I've seen far more offensive things than the use of the occasional rhetorical device!

[Laurentiu Cristofor]

The sensor of the Q, e.g., uses a back illuminated image sensor, i.e. other technology.

Agree.

But to clarify, in my mention of the compact@ISO100 vs the APS-C@ISO1600, what I had in mind were the Canon S3IS and the Pentax K10D which I purchased in the same year - 2007. The S3IS is the last compact camera I used, so that is my frame of reference for compact IQ. I happened to use it again a bit a year or two ago and I remember I was pleasantly surprised by its IQ, even though I was used with DSLR IQ by then.

[jstevewhite]

Display size

An aside...

The whole "equivalence" discussion reminds me of Polaroids. I would set up the shot with my Linhof and strobes, and put in a Polaroid back with type 59 and shoot a shot. Then I would show it to the art director, and explain "This is *kinda* what the image will look like, only the lighting will look better, the skin tones will look better, and the overall color and contrast will be much better. And that black will be black, too. And her eyes will be blue. In other words, it won't look very much like this at all, but here you go."

Only now, I don't HAVE to do that; so why WOULD I? Does anyone need to show an art director what the Q's images will look like, sorta, but not really?

[froeschle]

@jstevewhite
You know Pointilism? Pointillism - Wikipedia, the free encyclopedia
For a(n) (over)simplified comparison, imagine a picture made up of such points.
Furthermore, let us call the individual markings "photons".
The information content of an image then would be made up by a certain number of these.
Coded like that, it doesn't matter which physical size your recording or final image has.
You capture all the details, when you capture all the photons.
Now think of an image sensor duplicating this image.
What matters is reproducing all points - irrespective of the size of the sensor.
How many of the photons hit a certain area is unimportant, only the total number counts.
Then, "smaller size" or "magnification" is not the primary reason for a difference between systems - the total amount of light is.
So, exposure without mentioning sensor size is rather irrelevant for comparisons.
This should not be confusing - but clarifying relevant properties for cross-system analogies.
(Only) Equivalence here really provides "real world information".
Perhaps, we should discuss this - without insults and hostilities - in more detail in a separate thread, as this is not directly linked to "Pentax Q mirrorless: Specifications and lens details posted!"

The S3IS is the last compact camera I used, so that is my frame of reference for compact IQ. I happened to use it again a bit a year or two ago and I remember I was pleasantly surprised by its IQ, even though I was used with DSLR IQ by then.

This comparison also could be treated within this thread and we could mention why or why not additional differences exist.

[jstevewhite]

@jstevewhite
You know Pointilism? Pointillism - Wikipedia, the free encyclopedia
For a(n) (over)simplified comparison, imagine a picture made up of such points.
Furthermore, let us call the individual markings "photons".
The information content of an image then would be made up by a certain number of these.
Coded like that, it doesn't matter which physical size your recording or final image has.
You capture all the details, when you capture all the photons.
Now think of an image sensor duplicating this image.
What matters is reproducing all points - irrespective of the size of the sensor.
How many of the photons hit a certain area is unimportant, only the total number counts.
Then, "smaller size" or "magnification" is not the primary reason for a difference between systems - the total amount of light is.
So, exposure without mentioning sensor size is rather irrelevant for comparisons.
This should not be confusing - but clarifying relevant properties for cross-system analogies.
(Only) Equivalence here really provides "real world information".
Perhaps, we should discuss this - without insults and hostilities - in more detail in a separate thread, as this is not directly linked to "Pentax Q mirrorless: Specifications and lens details posted!"

This comparison also could be treated within this thread and we could mention why or why not additional differences exist.

I'm sorry if I've given the impression that I'm hostile, or if anyone feels insulted; it's not my intent. I do sometimes get somewhat animated

I completely understand what you're saying about photons and images, and I understand the claims of the equivalency analogy presented.

Basic physics doesn't seem to agree with you. Photons are small. Smaller than electrons, or atoms, or molecules, or sensors. Substrates are *always* resolution limited, for this fact if no other. You cannot continue to concentrate photons and continue to record detail, period. Sure, you can do the math, but the reality of things is that there is a lower limit - that comes far higher in the strata than molecules, I wager. If our substrate is resolution limited (as are digital sensors and film), then it is, indeed, the magnification that's the problem, because *sensor size is the problem*. Each technology improves the quality at the substrate, the recording media, but there are limits, and within a given technology, if you want greater IQ, all things being equal, you *must increase sensor size*. And if you CHANGE technology, equivalency ceases to be relevant even in the limited analogy it works in now.

Feel free to start a new thread and link it here, if you think it's warranted. I still think this is relevant here, as the claim was made that the IQ of the Q will be poor because it has a slow lens, which is inaccurate IMO, and that's what this discussion has been about. The equivalency analogy does not make the lens of the Q slow, and its IQ is limited by sensor size, NOT by lens speed, regardless of what the analogy of equivalency says.

[Class A]

I can put the same lens on both sensors. It's OBVIOUS that the same lens allows the SAME LIGHT through, yes?

Yes, but a 8.5mm f/1.9 (Q-system) and a 47mm f/1.9 (FF) are not the same lens.

Just google "What is a fast lens?" and everyone I can find uses it as I do, meaning "A lens is called fast because it allows a fast shutter speed".

I could present you with many examples where people use "fast" to reference lenses that produce shallow DOF. However, I have no interest in winning this argument this way. As froeschle suggested, let's not quibble over what "fast" means. IMO, in film days it primarily referred to the ability to use higher shutter speeds. Now that ISO values such as ISO 6400 are usable, even slow lenses can be used in combination with high shutter speeds. The relevance of a "fast" lens has therefore shifted to being able to produce shallow DOF.

If you keep the sensor format fixed (and for the sake of the argument the focal length as well) then a faster lens will produce shallower DOF. That's why it makes sense to use "fast" in this sense as well.

Since @ISO 100, the Q will use the same shutter speed as, say, the 43 f1.9 on my K-5, it is just as *fast*.

As froeschle nicely pointed out with his "pointilism" argument, using the same shutter speed doesn't mean a thing. You get the same exposure alright, but you get a lot less photons and the total amount of light matters, not that you can capture 1/6,000,000 of the photons of a larger sensor, if you only make the sensor small enough.

It's completely counter-productive to attempt to re-define f-stops and exposure and the like

No one is redefining those terms. The only thing I'm saying is that "same exposure" is not a useful thing to aspire to if you want the "same image quality".

It's NOT THE LENSES. It's the SENSORS.

Wrong.
You seemed to have preferred how Falk put things, so here are a few quotes from him (from the "Low noise benefit of FF vs APS-C equals ... zero" thread that I pointed you before):

It isn't the sensor which collects the photons, it is the lens!

To be explicit: the sensor size doesn't come into play here!

However, "your" f4 lens on FF is bigger and therefore, gathers more light.

Please continue that fight with Falk. The way I explain things obviously doesn't work for you.

Here's a concept that's missing from your equivalence equation: Final magnification.

It would really help if you read the sources I provided you with (see point 9). Of course, equivalence assumes the same output size and therefore factors in the final magnification.

And none of these require me to say things that will confuse people into thinking they have to use different settings to get the same exposure.

"confuse into thinking" -> "educate to understand"

...because they're only analogies that don't describe real-world photographic usage.

Wrong.

Basic physics doesn't seem to agree with you.

Basic physics does agree with froeschle. Equivalence is based on basic physics. It is not some esoteric principle. It is based on the notion of avoiding to compare apples with oranges. Once you accept that you have to compare formats of different size on the basis of producing equivalent images, all the rest follows with mathematical rigour, using basic physics.

if you want greater IQ, all things being equal, you *must increase sensor size*.

No, you must increase the number of photons used to create the image. Are you saying there will be non-linear superimposition of photons on a sensor the size of that used in the Q-system? With all due respect, this is nonsense.

And if you CHANGE technology, equivalency ceases to be relevant even in the limited analogy it works in now.

With all due respect: That doesn't even begin to make sense.

[Class A]

The information content of an image then would be made up by a certain number of these.
Coded like that, it doesn't matter which physical size your recording or final image has.

Exactly.

How many of the photons hit a certain area is unimportant, only the total number counts.

Exactly.

Then, "smaller size" or "magnification" is not the primary reason for a difference between systems - the total amount of light is.

Exactly.
The significance of magnification depends on the initial resolution, so cannot be regarded in isolation. The argument that sensors can become too small to capture a certain amount of resolution doesn't make sense at the scale we are talking about.

So, exposure without mentioning sensor size is rather irrelevant for comparisons.

Exactly.

[jstevewhite]

Yes, but a 8.5mm f/1.9 (Q-system) and a 47mm f/1.9 (FF) are not the same lens.

Agreed, they are not the same lens. This much, we're square on.

The relevance of a "fast" lens has therefore shifted to being able to produce shallow DOF.

Agree, to disagree

Basic physics does agree with froeschle. Equivalence is based on basic physics. It is not some esoteric principle. It is based on the notion of avoiding to compare apples with oranges. Once you accept that you have to compare formats of different size on the basis of producing equivalent images, all the rest follows with mathematical rigour, using basic physics.

I understand completely the mathematical model that Falk is talking about. I even understand why you - and, as you quote him, he - make counter-factual claims. I'm just pointing out that the claims *are* counter-factual. As I explained, I can put the same lens on FF and on APS-c (and yes, on the Q), and use the same exposure settings to obtain the same exposure. If this is in fact true, then it cannot be the case that the lens lets through more light - it's the same lens; it can ONLY be that the sensor captures less of the light that it does pass.

That is what I'd like for you to explain, in context. How the same 8.5mm lens can pass a certain amount of light on ONE format, but a different amount when mounted on another.

No, you must increase the number of photons used to create the image. Are you saying there will be non-linear superimposition of photons on a sensor the size of that used in the Q-system? With all due respect, this is nonsense.

With all due respect: That doesn't even begin to make sense.

Every sensor technology - ever sensor - has a "full well" capacity per pixel. The ideal exposure places the whitest white at or near that full-well capacity. The higher the full-well capacity, the higher the signal-to-noise ratio, and the higher the dynamic range. Smaller pixels have smaller full-well capacities. If you did concentrate all the light that falls on the larger sensor onto the smaller sensor of, say, the Q, you'd have nearly every pixel pegged to max, because they have lower full well capacities. You would have to change to a different (currently unknown) technology to make this happen. In the *real world*, the full well capacity of pixels is what limits sensors, not their lenses. Even if I did build that f0.3 lens, the nominal ISO of the sensor is what it is, and if I followed your guideline, I would overexpose it horribly.

So while I understand the mathematical concept expressed, it doesn't fit the real world. That is, it only 'works' in one direction (that is, the large sensor can emulate the smaller sensor by increasing ISO; but the small sensor cannot emulate the larger one by reducing ISO beyond its base level), and for certain situations, whereas the standard definition of exposure works with all lenses and formats consistently and accurately. Additionally, the standard model of exposure isn't faced with absurdities like claiming that I could, say, focus all the photons used by a 6x6cm back onto one molecule and get the same image quality as the 6x6cm image - that is, the absolute physical limitations of sensor manufacture (which I suspect are considerably larger than one molecule anyway).

So it's fine to say something like "If you want to see something that looks kinda like what you might get from the Q, possibly, try setting your APS-c camera to ISO1600 and put on a 30mm lens and set the aperture to 7.1 and shoot some pictures. It might look something a little bit like that." But to make claims like "The lens passes less light because it's attached to a camera with a smaller sensor" is just absurd.

Here's a good article discussing the physical limitation of sensor size and pixel size:

http://www.clarkvision.com/articles/digital.signal.to.noise/

And more:

http://www.clarkvision.com/articles/does.pixel.size.matter/

[unfocused]

Dear God... did film guys get bogged down in this trivia when deciding which format to use - 828, 135, half-frame, 126, 127, 110, Minox, etc? No, they just went out and shot their pictures.

[jstevewhite]

Dear God... did film guys get bogged down in this trivia when deciding which format to use - 828, 135, half-frame, 126, 127, 110, Minox, etc? No, they just went out and shot their pictures.

LOL! Actually, there were constant debates in camera clubs (which were the only real analog of forums back then), which usually ended with the 35mm guy saying "Well, it's good enough for everything *I* do, and I would like to see you shoot a baseball game with that Speed Graphic!" and the medium format guy saying "You can't seriously consider that 'acceptable quality', right?" and the large format guy looking at both of them and saying "By the time you enlarge that thing it's going to look like it was drawn with crayons."

And the Leica guy sitting on the tall barstool in the corner, fondling his M3 and smiling silently.

The exact same discussion applies to film, incidentally, and you would, from time to time, hear people talking about things like "Tri-X in my Rolleiflex TLR looks as fine-grained as Plus-X in your 35mm camera." There were certainly discussions about DOF with various formats, but they went like this: "Oh, you're shooting 4x5? You're going to need at least F22, maybe f32, to get that DOF. Or you'll have to swing the standards."

It was very rare for anyone to ask how to get LESS DOF back in the day (meaning, prior to 2000 or so). But you would hear "Oh, you need more DOF? Try a wider angle lens."

[kadajawi]

Great. I'm with you. I think we just have a different definition for the word, "fast." This seems to be at the heart of a large portion of the discussion in this thread. Is there some sort of official reference for this term, or is it just a convention? If it's just a convention, then I think the definition I am using is more conventional than the one you are using.

I don't think the conventional definition of "fast" includes depth of field. DOF is a corollary, but it's not part of the commonly understood definition. In other words, if the conventional understanding of a fast lens is that with a fast lens you can use a fast shutter speed at a given ISO in given lighting conditions to get the same exposure because you can use a smaller f-stop value, then by this convention, the 8.5mm f/1.9 lens of the Q is faster than a 31mm f/7 lens.

What I'm saying is it might be impeding understanding to use the terms slow and fast when talking about depth of field (I know it impeded my understanding). Again, this might be due to my lack of experience. If the conventional understanding is that a fast lens has a shallow depth of field and lets in a lot of light, then I'm off base.

I agree, to me a fast lens is one that lets a lot light in (short shutter speeds). Wikipedia agrees. Lens speed - Wikipedia, the free encyclopedia Of course a fast lens _usually_ results in shallow DoF, but that is not necessarily the case (as in a fast fish eye lens). Thus considering the DoF when calculating equivalence is IMHO rather problematic, as the lens in that case lets in less light.

[unfocused]

LOL! Actually, there were constant debates in camera clubs (which were the only real analog of forums back then), which usually ended with the 35mm guy saying "Well, it's good enough for everything *I* do, and I would like to see you shoot a baseball game with that Speed Graphic!" and the medium format guy saying "You can't seriously consider that 'acceptable quality', right?" and the large format guy looking at both of them and saying "By the time you enlarge that thing it's going to look like it was drawn with crayons."

And the Leica guy sitting on the tall barstool in the corner, fondling his M3 and smiling silently.

The exact same discussion applies to film, incidentally, and you would, from time to time, hear people talking about things like "Tri-X in my Rolleiflex TLR looks as fine-grained as Plus-X in your 35mm camera." There were certainly discussions about DOF with various formats, but they went like this: "Oh, you're shooting 4x5? You're going to need at least F22, maybe f32, to get that DOF. Or you'll have to swing the standards."

It was very rare for anyone to ask how to get LESS DOF back in the day (meaning, prior to 2000 or so). But you would hear "Oh, you need more DOF? Try a wider angle lens."

Ha, you bring back memories! Yes I spent many years in photo clubs and, yes, the techno-freaks loved their arguments. "Bullshit baffles brains", remember. The real photographers simply got on with the art of photography, often making great pictures with the simplest "paintbrush". The acid test was their scores in the slide and print competitions/exhibitions.